452 CHEMICAL ENGINEERING objects in the furnace can be distinguished but only on close observation, and if much of the detail is lost, after they have been in the furnace some time, it is not likely that the temperature measurement will be seriously in error. If in error at all, the observed temperature will be too high when the furnace walls are brighter than the material being heat treated and too low when the walls are less bright. This latter condition is possible if the heat supply is variable or if it is shut off and the furnace allowed to cool. That a steel ingot placed in a heated furnace may appear much hotter than it really is, is a fact not always appreciated. The surface of the ingot appears hot because it reflects the bright light from the walls of the furnace. Of course in comparison to the much greater brightness of the walls, the cold ingot appears black, but this is due to the intense contrast. If the ingot is viewed alone when the direct radiation from the furnace is screened from the eye, it also is bright. Thus when an optical pyrometer is sighted on an ingot in the furnace, part of the light reaching the instrument comes from the side walls and is reflected by the surface of the ingot. Iron oxide reflects about the least amount of any material met with in metallurgical practice. Its emissivity is approximately 0.95; hence its reflection coefficient is 0.05. Suppose an iron ingot at room temperature were suddenly placed in a furnace at 1,200°C. Although the surface of the ingot is cold it reflects 5 per cent of the light falling upon it from the hot side walls. This 5 per cent of reflected radiation gives the ingot the appearance of an object at 950°C. The measurement by the optical pyrometer accordingly would be 950°C., although actually the ingot is at room temperature. The higher the reflecting power of the material the greater the observed temperature under the above conditions. Thus cold platinum would appear to be at about 1,160° or at almost the same temperature as that of the furnace. One method for reducing the error due to reflected radiation is to view the object through a large open door on a surface parallel to the opening. If the door is large enough the surface of the ingot thus loses a great portion of the furnace wall contributing the reflected light. The ingot reflects the image of the opening left by the door and this opening is not radiating. It must be pointed out that the reflection from any material being heat treated is diffuse like that from a matt surface so that an opening in a furnace wall is not truly imaged as would be the case with a polished mirror surface. Stray reflection is, however, greatly minimized even with matt surfaces by sighting through a large door on a surface parallel to the opening. When the material has attained the temperature of the side walls it is of course not desirable to open a large door since then the opening affects the black-body conditions. The pyrometer should be sighted through a small peep hole as soon as approximate temperature uniformity is obtained. A more satisfactory method for reducing the stray reflections than by opening a large door is to sight into a deep wedge-shaped cavity or hole made in the metal being heat treated. If this cavity is deep enough, very little radiation from the side walls can be reflected from it. If such a hole cannot be made conveniently, a length of iron pipe closed at one end, or a porcelain tube, may be placed on the material and so aligned that the pyrometer may be sighted through a peep hole directly into the bottom of the tube. The effect of reflected light is very noticeable in an empty coke oven. The reflection coefficient of the brick walls is comparatively high. If a perfect mirror were placed parallel to a glowing surface the mirror would appear as hot as the surface. This effect of reflection takes place in a coke oven so that both walls appear of approximately equal brightness even though they may differ considerably in temperature. Frequently a patch of the wall on one side becomes coated with a layer of coke. Since the coke has a higher emissive power than brick this patch appears much hotter.